Further reduce extended attribute reads/writes

More refactoring towards reducing the number of extended attribute
reads/writes to the file system. See BZ 868120:
    http://bugzilla.redhat.com/show_bug.cgi?id=868120

Basically the redundant routines restore_object, restore_account and
restore_container have been collapsed to one routine, restore_metadata, which
will only write out metadata if the new metadata is different from what was
originally read.

Along with these changes come a set of unit tests for all the functions
related to the routines that are in this module in the call tree where
restore_metadata is invoked.

Change-Id: I957ee2f8646cbe6df4d4420d3bdfb1f6ac62bdd2
BUG: 868120
Signed-off-by: Peter Portante <peter.portante@redhat.com>
Reviewed-on: http://review.gluster.org/4111
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Tested-by: Anand Avati <avati@redhat.com>

3 files changed, 279 insertions, 38 deletions

diff --git a/swift/1.4.8/.unittests b/swift/1.4.8/.unittests
new file mode 100755
index 000000000..113e5dd1f
--- /dev/null
+++ b/swift/1.4.8/.unittests
@@ -0,0 +1,6 @@
+#!/bin/bash
+
+cd test/unit
+nosetests --exe --with-coverage --cover-package plugins --cover-erase $@
+rm -f .coverage
+cd -
diff --git a/swift/1.4.8/plugins/utils.py b/swift/1.4.8/plugins/utils.py
index ef5f9d62f..289420cb9 100644
--- a/swift/1.4.8/plugins/utils.py
+++ b/swift/1.4.8/plugins/utils.py
@@ -98,7 +98,6 @@ def do_makedirs(path):
             raise
     return True
 
-
 def do_listdir(path):
     try:
         buf = os.listdir(path)
@@ -278,7 +277,6 @@ def dir_empty(path):
         if not os.path.exists(path):
             return True
 
-
 def get_device_from_account(account):
     if account.startswith(RESELLER_PREFIX):
         device = account.replace(RESELLER_PREFIX, '', 1)
@@ -517,7 +515,6 @@ def get_account_details_from_memcache(acc_path, memcache=None):
         container_count = memcache.get(strip_obj_storage_path(acc_path)+'_container_count')
         return container_list, container_count
 
-
 def get_account_details(acc_path, memcache=None):
     """
     Return container_list and container_count.
@@ -527,8 +524,6 @@ def get_account_details(acc_path, memcache=None):
     else:
         return get_account_details_from_fs(acc_path, memcache)
 
-
-
 def get_etag(path):
     etag = None
     if os.path.exists(path):
@@ -613,45 +608,28 @@ def get_account_metadata(acc_path, memcache=None):
                 X_CONTAINER_COUNT: container_count}
     return _add_timestamp(metadata)
 
-def restore_object(obj_path, metadata):
-    meta = read_metadata(obj_path)
-    if meta:
-        meta.update(metadata)
-        write_metadata(obj_path, meta)
+def restore_metadata(path, metadata):
+    meta_orig = read_metadata(path)
+    if meta_orig:
+        meta_new = meta_orig.copy()
+        meta_new.update(metadata)
     else:
-        write_metadata(obj_path, metadata)
-
-def restore_container(cont_path, metadata):
-    meta = read_metadata(cont_path)
-    if meta:
-        meta.update(metadata)
-        write_metadata(cont_path, meta)
-    else:
-        write_metadata(cont_path, metadata)
-
-def restore_account(acc_path, metadata):
-    meta = read_metadata(acc_path)
-    if meta:
-        meta.update(metadata)
-        write_metadata(acc_path, meta)
-    else:
-        write_metadata(acc_path, metadata)
+        meta_new = metadata
+    if meta_orig != meta_new:
+        write_metadata(path, meta_new)
+    return meta_new
 
 def create_object_metadata(obj_path):
-    meta = get_object_metadata(obj_path)
-    restore_object(obj_path, meta)
-    return meta
+    metadata = get_object_metadata(obj_path)
+    return restore_metadata(obj_path, metadata)
 
 def create_container_metadata(cont_path, memcache=None):
-    meta = get_container_metadata(cont_path, memcache)
-    restore_container(cont_path, meta)
-    return meta
+    metadata = get_container_metadata(cont_path, memcache)
+    return restore_metadata(cont_path, metadata)
 
 def create_account_metadata(acc_path, memcache=None):
-    meta = get_account_metadata(acc_path, memcache)
-    restore_account(acc_path, meta)
-    return meta
-
+    metadata = get_account_metadata(acc_path, memcache)
+    return restore_metadata(acc_path, metadata)
 
 def check_account_exists(account, fs_object):
     if account not in get_account_list(fs_object):
@@ -666,7 +644,6 @@ def get_account_list(fs_object):
         account_list = fs_object.get_export_list()
     return account_list
 
-
 def get_account_id(account):
     return RESELLER_PREFIX + md5(account + HASH_PATH_SUFFIX).hexdigest()
 
diff --git a/swift/1.4.8/test/unit/plugins/test_utils.py b/swift/1.4.8/test/unit/plugins/test_utils.py
index daecc425c..944d11892 100644
--- a/swift/1.4.8/test/unit/plugins/test_utils.py
+++ b/swift/1.4.8/test/unit/plugins/test_utils.py
@@ -15,10 +15,14 @@
 
 """ Tests for plugins.utils """
 
+import os
 import unittest
 import errno
 import xattr
 import cPickle as pickle
+import tempfile
+import hashlib
+from swift.common.utils import normalize_timestamp
 from collections import defaultdict
 from plugins import utils
 
@@ -229,3 +233,257 @@ class TestUtils(unittest.TestCase):
         assert res_d == {}
         assert _xattr_op_cnt['get'] == 3, "%r" % _xattr_op_cnt
         assert len(_xattrs.keys()) == 0, "Expected 0 keys, found %d" % len(_xattrs.keys())
+
+    def test_restore_metadata_none(self):
+        # No initial metadata
+        path = "/tmp/foo/i"
+        res_d = utils.restore_metadata(path, { 'b': 'y' })
+        expected_d = { 'b': 'y' }
+        assert res_d == expected_d, "Expected %r, result %r" % (expected_d, res_d)
+        assert _xattr_op_cnt['get'] == 1, "%r" % _xattr_op_cnt
+        assert _xattr_op_cnt['set'] == 1, "%r" % _xattr_op_cnt
+
+    def test_restore_metadata(self):
+        # Initial metadata
+        path = "/tmp/foo/i"
+        initial_d = { 'a': 'z' }
+        xkey = _xkey(path, utils.METADATA_KEY)
+        _xattrs[xkey] = pickle.dumps(initial_d, utils.PICKLE_PROTOCOL)
+        res_d = utils.restore_metadata(path, { 'b': 'y' })
+        expected_d = { 'a': 'z', 'b': 'y' }
+        assert res_d == expected_d, "Expected %r, result %r" % (expected_d, res_d)
+        assert _xattr_op_cnt['get'] == 1, "%r" % _xattr_op_cnt
+        assert _xattr_op_cnt['set'] == 1, "%r" % _xattr_op_cnt
+
+    def test_restore_metadata_nochange(self):
+        # Initial metadata but no changes
+        path = "/tmp/foo/i"
+        initial_d = { 'a': 'z' }
+        xkey = _xkey(path, utils.METADATA_KEY)
+        _xattrs[xkey] = pickle.dumps(initial_d, utils.PICKLE_PROTOCOL)
+        res_d = utils.restore_metadata(path, {})
+        expected_d = { 'a': 'z' }
+        assert res_d == expected_d, "Expected %r, result %r" % (expected_d, res_d)
+        assert _xattr_op_cnt['get'] == 1, "%r" % _xattr_op_cnt
+        assert _xattr_op_cnt['set'] == 0, "%r" % _xattr_op_cnt
+
+    def test_add_timestamp_empty(self):
+        orig = {}
+        res = utils._add_timestamp(orig)
+        assert res == {}
+
+    def test_add_timestamp_none(self):
+        orig = { 'a': 1, 'b': 2, 'c': 3 }
+        exp = { 'a': (1, 0), 'b': (2, 0), 'c': (3, 0) }
+        res = utils._add_timestamp(orig)
+        assert res == exp
+
+    def test_add_timestamp_mixed(self):
+        orig = { 'a': 1, 'b': (2, 1), 'c': 3 }
+        exp = { 'a': (1, 0), 'b': (2, 1), 'c': (3, 0) }
+        res = utils._add_timestamp(orig)
+        assert res == exp
+
+    def test_add_timestamp_all(self):
+        orig = { 'a': (1, 0), 'b': (2, 1), 'c': (3, 0) }
+        res = utils._add_timestamp(orig)
+        assert res == orig
+
+    def test_get_etag_empty(self):
+        tf = tempfile.NamedTemporaryFile()
+        hd = utils.get_etag(tf.name)
+        assert hd == hashlib.md5().hexdigest()
+
+    def test_get_etag(self):
+        tf = tempfile.NamedTemporaryFile()
+        tf.file.write('123' * utils.CHUNK_SIZE)
+        tf.file.flush()
+        hd = utils.get_etag(tf.name)
+        tf.file.seek(0)
+        md5 = hashlib.md5()
+        while True:
+            chunk = tf.file.read(utils.CHUNK_SIZE)
+            if not chunk:
+                break
+            md5.update(chunk)
+        assert hd == md5.hexdigest()
+
+    def test_get_object_metadata_dne(self):
+        md = utils.get_object_metadata("/tmp/doesNotEx1st")
+        assert md == {}
+
+    obj_keys = (utils.X_TIMESTAMP, utils.X_CONTENT_TYPE, utils.X_ETAG,
+                utils.X_CONTENT_LENGTH, utils.X_TYPE, utils.X_OBJECT_TYPE)
+
+    def test_get_object_metadata_file(self):
+        tf = tempfile.NamedTemporaryFile()
+        tf.file.write('123'); tf.file.flush()
+        md = utils.get_object_metadata(tf.name)
+        for key in self.obj_keys:
+            assert key in md, "Expected key %s in %r" % (key, md)
+        assert md[utils.X_TYPE] == utils.OBJECT
+        assert md[utils.X_OBJECT_TYPE] == utils.FILE
+        assert md[utils.X_CONTENT_TYPE] == utils.FILE_TYPE
+        assert md[utils.X_CONTENT_LENGTH] == os.path.getsize(tf.name)
+        assert md[utils.X_TIMESTAMP] == normalize_timestamp(os.path.getctime(tf.name))
+        assert md[utils.X_ETAG] == utils.get_etag(tf.name)
+
+    def test_get_object_metadata_dir(self):
+        td = tempfile.mkdtemp()
+        try:
+            md = utils.get_object_metadata(td)
+            for key in self.obj_keys:
+                assert key in md, "Expected key %s in %r" % (key, md)
+            assert md[utils.X_TYPE] == utils.OBJECT
+            assert md[utils.X_OBJECT_TYPE] == utils.DIR
+            assert md[utils.X_CONTENT_TYPE] == utils.DIR_TYPE
+            assert md[utils.X_CONTENT_LENGTH] == 0
+            assert md[utils.X_TIMESTAMP] == normalize_timestamp(os.path.getctime(td))
+            assert md[utils.X_ETAG] == hashlib.md5().hexdigest()
+        finally:
+            os.rmdir(td)
+
+    def test_create_object_metadata_file(self):
+        tf = tempfile.NamedTemporaryFile()
+        tf.file.write('4567'); tf.file.flush()
+        r_md = utils.create_object_metadata(tf.name)
+
+        xkey = _xkey(tf.name, utils.METADATA_KEY)
+        assert len(_xattrs.keys()) == 1
+        assert xkey in _xattrs
+        assert _xattr_op_cnt['get'] == 1
+        assert _xattr_op_cnt['set'] == 1
+        md = pickle.loads(_xattrs[xkey])
+        assert r_md == md
+
+        for key in self.obj_keys:
+            assert key in md, "Expected key %s in %r" % (key, md)
+        assert md[utils.X_TYPE] == utils.OBJECT
+        assert md[utils.X_OBJECT_TYPE] == utils.FILE
+        assert md[utils.X_CONTENT_TYPE] == utils.FILE_TYPE
+        assert md[utils.X_CONTENT_LENGTH] == os.path.getsize(tf.name)
+        assert md[utils.X_TIMESTAMP] == normalize_timestamp(os.path.getctime(tf.name))
+        assert md[utils.X_ETAG] == utils.get_etag(tf.name)
+
+    def test_create_object_metadata_dir(self):
+        td = tempfile.mkdtemp()
+        try:
+            r_md = utils.create_object_metadata(td)
+
+            xkey = _xkey(td, utils.METADATA_KEY)
+            assert len(_xattrs.keys()) == 1
+            assert xkey in _xattrs
+            assert _xattr_op_cnt['get'] == 1
+            assert _xattr_op_cnt['set'] == 1
+            md = pickle.loads(_xattrs[xkey])
+            assert r_md == md
+
+            for key in self.obj_keys:
+                assert key in md, "Expected key %s in %r" % (key, md)
+            assert md[utils.X_TYPE] == utils.OBJECT
+            assert md[utils.X_OBJECT_TYPE] == utils.DIR
+            assert md[utils.X_CONTENT_TYPE] == utils.DIR_TYPE
+            assert md[utils.X_CONTENT_LENGTH] == 0
+            assert md[utils.X_TIMESTAMP] == normalize_timestamp(os.path.getctime(td))
+            assert md[utils.X_ETAG] == hashlib.md5().hexdigest()
+        finally:
+            os.rmdir(td)
+
+    def test_get_container_metadata(self):
+        def _mock_get_container_details(path, memcache=None):
+            o_list = [ 'a', 'b', 'c' ]
+            o_count = 3
+            b_used = 47
+            return o_list, o_count, b_used
+        td = tempfile.mkdtemp()
+        orig_gcd = utils.get_container_details
+        utils.get_container_details = _mock_get_container_details
+        try:
+            exp_md = {
+                utils.X_TYPE: (utils.CONTAINER, 0),
+                utils.X_TIMESTAMP: (normalize_timestamp(os.path.getctime(td)), 0),
+                utils.X_PUT_TIMESTAMP: (normalize_timestamp(os.path.getmtime(td)), 0),
+                utils.X_OBJECTS_COUNT: (3, 0),
+                utils.X_BYTES_USED: (47, 0),
+                }
+            md = utils.get_container_metadata(td)
+            assert md == exp_md
+        finally:
+            utils.get_container_details = orig_gcd
+            os.rmdir(td)
+
+    def test_get_account_metadata(self):
+        def _mock_get_account_details(path, memcache=None):
+            c_list = [ '123', 'abc' ]
+            c_count = 2
+            return c_list, c_count
+        td = tempfile.mkdtemp()
+        orig_gad = utils.get_account_details
+        utils.get_account_details = _mock_get_account_details
+        try:
+            exp_md = {
+                utils.X_TYPE: (utils.ACCOUNT, 0),
+                utils.X_TIMESTAMP: (normalize_timestamp(os.path.getctime(td)), 0),
+                utils.X_PUT_TIMESTAMP: (normalize_timestamp(os.path.getmtime(td)), 0),
+                utils.X_OBJECTS_COUNT: (0, 0),
+                utils.X_BYTES_USED: (0, 0),
+                utils.X_CONTAINER_COUNT: (2, 0),
+                }
+            md = utils.get_account_metadata(td)
+            assert md == exp_md
+        finally:
+            utils.get_account_details = orig_gad
+            os.rmdir(td)
+
+    cont_keys = [utils.X_TYPE, utils.X_TIMESTAMP, utils.X_PUT_TIMESTAMP,
+                 utils.X_OBJECTS_COUNT, utils.X_BYTES_USED]
+
+    def test_create_container_metadata(self):
+        td = tempfile.mkdtemp()
+        try:
+            r_md = utils.create_container_metadata(td)
+
+            xkey = _xkey(td, utils.METADATA_KEY)
+            assert len(_xattrs.keys()) == 1
+            assert xkey in _xattrs
+            assert _xattr_op_cnt['get'] == 1
+            assert _xattr_op_cnt['set'] == 1
+            md = pickle.loads(_xattrs[xkey])
+            assert r_md == md
+
+            for key in self.cont_keys:
+                assert key in md, "Expected key %s in %r" % (key, md)
+            assert md[utils.X_TYPE] == (utils.CONTAINER, 0)
+            assert md[utils.X_TIMESTAMP] == (normalize_timestamp(os.path.getctime(td)), 0)
+            assert md[utils.X_PUT_TIMESTAMP] == (normalize_timestamp(os.path.getmtime(td)), 0)
+            assert md[utils.X_OBJECTS_COUNT] == (0, 0)
+            assert md[utils.X_BYTES_USED] == (0, 0)
+        finally:
+            os.rmdir(td)
+
+    acct_keys = [val for val in cont_keys]
+    acct_keys.append(utils.X_CONTAINER_COUNT)
+
+    def test_create_account_metadata(self):
+        td = tempfile.mkdtemp()
+        try:
+            r_md = utils.create_account_metadata(td)
+
+            xkey = _xkey(td, utils.METADATA_KEY)
+            assert len(_xattrs.keys()) == 1
+            assert xkey in _xattrs
+            assert _xattr_op_cnt['get'] == 1
+            assert _xattr_op_cnt['set'] == 1
+            md = pickle.loads(_xattrs[xkey])
+            assert r_md == md
+
+            for key in self.acct_keys:
+                assert key in md, "Expected key %s in %r" % (key, md)
+            assert md[utils.X_TYPE] == (utils.ACCOUNT, 0)
+            assert md[utils.X_TIMESTAMP] == (normalize_timestamp(os.path.getctime(td)), 0)
+            assert md[utils.X_PUT_TIMESTAMP] == (normalize_timestamp(os.path.getmtime(td)), 0)
+            assert md[utils.X_OBJECTS_COUNT] == (0, 0)
+            assert md[utils.X_BYTES_USED] == (0, 0)
+            assert md[utils.X_CONTAINER_COUNT] == (0, 0)
+        finally:
+            os.rmdir(td)